pi-code 1.0.55 → 1.0.56

package/README.md

@@ -13,7 +13,7 @@
 
 Claude Code experience for the [pi](https://pi.dev) coding agent, in one package. Point pi at a project that already has a `.claude/` directory and it reads your existing config: rules, commands, skills, hooks, output styles, MCP servers, and agents. It also adds the Claude Code features pi lacks: a todo overlay, checkpoints, memory, web search, subagents, and goals.
 
-What a repository ships is treated as untrusted until you approve it: project MCP servers, hooks, agents, rules, output styles, commands and skills load only once you say yes.
+What a repository ships is treated as untrusted until you approve it: project MCP servers, hooks, agents, rules, output styles, commands and skills load only once you say yes. A headless run (`pi -p`) cannot ask, so an undecided project loads none of them there, which is stricter than Claude, where a headless run uses them without showing the dialog.
 
 ![pi-code demo](demos/hero.gif)
 
@@ -45,7 +45,7 @@ Each topic links to its own doc with the full contract and any divergences from
 - **[MCP servers](docs/mcp.md)** — every Claude config scope, all four transports, OAuth, managed policy, timeouts, prompts, and resources.
 - **[Custom slash commands](docs/commands.md)** — `.claude/commands` with arguments, bash spans, `@file` inlining, frontmatter, and model invocation.
 - **[Skills](docs/skills.md)** — `.claude/skills` discovery plus the same dynamic content commands get.
-- **[Subagents / Task](docs/subagents.md)** — built-in and custom agents, background runs, per-agent memory, worktree isolation.
+- **[Subagents](docs/subagents.md)** — built-in and custom agents, background runs, per-agent memory, worktree isolation.
 - **[CLAUDE.md, @imports, and rules](docs/claude-md.md)** — the context files and path-scoped rules pi does not load natively.
 - **[Settings `env`](docs/settings-env.md)** — env blocks from every settings scope, exported with Claude's precedence.
 - **[Output styles](docs/output-styles.md)** — replace semantics, bundled built-ins, `/output-style`.

package/extensions/hooks/runners.ts

@@ -224,7 +224,11 @@ export async function runHttpHook(hook: { type?: string; command: string; url?:
     }
     return { code: 0, stdout: body, stderr: '', timedOut: false }
   } catch (error) {
-    return { code: 1, stdout: '', stderr: error instanceof Error ? error.message : String(error), timedOut: false }
+    // A timeout is the abort that AbortSignal.timeout raises. Mark it as one so a gated
+    // event fails closed on it, exactly as a command hook that ran out of time does: the
+    // hook never answered, whichever transport it used.
+    const name = error instanceof Error ? error.name : ''
+    return { code: 1, stdout: '', stderr: error instanceof Error ? error.message : String(error), timedOut: name === 'TimeoutError' || name === 'AbortError' }
   }
 }
 
@@ -319,7 +323,8 @@ export async function runMcpToolHook(hook: HookCommand, payload: unknown, timeou
   const input = hook.input && typeof hook.input === 'object' ? (substituteInputPaths(hook.input, payload) as Record<string, unknown>) : {}
   let timer: ReturnType<typeof setTimeout> | undefined
   const deadline = new Promise<HookRunResult>((resolve) => {
-    timer = setTimeout(() => resolve({ code: 1, stdout: '', stderr: `mcp_tool hook timed out after ${timeoutMs}ms`, timedOut: false }), timeoutMs)
+    // Marked as a timeout so a gated event fails closed on it, like a command hook.
+    timer = setTimeout(() => resolve({ code: 1, stdout: '', stderr: `mcp_tool hook timed out after ${timeoutMs}ms`, timedOut: true }), timeoutMs)
   })
   const call = callMcpTool(hook.server, hook.tool, input)
     .then((result): HookRunResult => ({ code: result.isError ? 1 : 0, stdout: result.text, stderr: '', timedOut: false }))

package/extensions/subagent/README.md

@@ -39,7 +39,7 @@ This tool executes a separate `pi` subprocess with a delegated system prompt and
 
 **Default behavior:** Loads the bundled builtin agents (Explore, Plan, general-purpose) plus **user-level agents** from `~/.claude/agents` and `~/.pi/agent/agents`. A user or project agent with the same name overrides a builtin. Discovered agents and their descriptions are listed in the system prompt each turn, so the model can pick one itself; project agent descriptions appear only once the project is approved.
 
-To enable project-local agents (`.claude/agents`, `.pi/agents`), pass `agentScope: "both"` (or `"project"`). Only do this for repositories you trust.
+Project-local agents (`.claude/agents`, `.pi/agents`) load once the project is approved: a default call resolves `agentScope` to `"both"` for an approved project and `"user"` otherwise, and an explicit `agentScope` still narrows or widens it. Every invocation passes the project-agent gate either way.
 
 When running interactively, the tool prompts for confirmation before running project-local agents. `confirmProjectAgents: false` skips that prompt for a project you have already approved; an unapproved project is still asked about.
 
@@ -134,9 +134,9 @@ enforces per call.
 
 **Locations:**
 - `~/.claude/agents/*.md`, `~/.pi/agent/agents/*.md` - User-level (always loaded; `~/.pi` wins a name conflict)
-- `.claude/agents/*.md`, `.pi/agents/*.md` - Project-level (only with `agentScope: "project"` or `"both"`; `.pi` wins a name conflict)
+- `.claude/agents/*.md`, `.pi/agents/*.md` - Project-level (an approved project, or an explicit `agentScope` of `"project"`/`"both"`; `.pi` wins a name conflict)
 
-Project agents override user agents with the same name when `agentScope: "both"`.
+Project agents override user agents with the same name when both scopes are in play.
 
 ## Builtin Agents
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-code",
-  "version": "1.0.55",
+  "version": "1.0.56",
   "description": "Claude Code experience for the pi coding agent: reads your .claude config (rules, commands, skills, hooks, output styles, MCP servers, agents) and adds todo, checkpoints, memory, web, subagents, and goals",
   "keywords": [
     "pi",